Stats 1.4-1.6 Vocab

stratified sample

divides the population into separate groups (strata) then selects a simple random sample from each stratum

• individuals within each stratum should be homogeneous with respect to the variable of interest

• this method guarantees that each stratum is represented in the sample

cluster sample

divides the population into a large number of clusters (city blocks), then a simple random cluster is selected and ALL individuals in the selected clusters are included in the sample

• cluster sampling is less expensive per observation than simple random sampling

systematic sample

obtained by selecting every kth individual from a population, the first individual selected corresponds to a random number between 1-k

• does not require a list (sampling frame ex. polling)

convenience sample

samples in which individuals are easily obtained and not based on randomness (self-selected ex. amazon reviews)

• also called “voluntary response samples”

bias

if a sample has bias that is systematically different from the population with respect to the variable, leads to under-estimation or an over-estimation of the parameter

sampling bias

results when the technique used to obtain the samples favor one part of the population over another

• poor sampling ex. undercoverage, non-random sampling

nonresponse bias

exists when individuals selected to be in sample DO NOT respond to the survey and have different opinions from those who do respond

response bias

exists when answers on a survey DO NOT reflect the true opinions of the respondent

• lying or leading wordy questions

randomization

randomly assign subjects to treatment to eliminate the bias that may result if subjects are assigned to balance group variables

• lurking variables can occur

blinding

treatment groups are treated equally as possible

single-blind

subjects DO NOT know which treatment is being recieved

double-blind

neither subject or researcher are aware of what treatment the subjects are receiving

replication

each treatment is applied to more than one experimental unit, using more than one experimental unit for each treatment ensuring the effect of a treatment

completely randomized design

each experimental unit is randomly assigned to one treatment group

can medical studies use random subjects?

no, medical studies must use subjects that have the medical issue for the designed treatment that is being tested